Drop the GitHub-release version comparison entirely. We install slirp4netns
from apt regardless, so comparing against the GitHub-latest tag only produced
a perpetual 'outdated' loop and a no-op re-install. apt-get install -y is
already idempotent, so run it unconditionally and report the resulting
version.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
slirp4netns --version prints multiple lines (version, commit, libslirp,
SLIRP_CONFIG_VERSION_MAX). The old 'awk {print $2}' ran on every line and
also picked the literal word 'version' from line 1, producing a multi-line
blob that leaked into the 'is outdated' notice. Read only the first line and
take field 3 (the actual number), strip the leading v from the GitHub tag so
the comparison is meaningful, and skip the check if the tag fetch fails.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Initial rootless setup ran 'systemctl --user start docker' immediately
after install, but the rootless net/port-driver override.conf (and the
daemon-reload that loads it) aren't written until further down. So the
first start always failed — 'Job for docker.service failed' plus a
spurious '✗ Error Setting up Rootless' in the error report — even though
the later 'systemctl --user restart docker' brought the daemon up fine
once the override was in place.
Drop the premature start from the install step (keep install + enable);
the restart after the override is written is now the first real start.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The Services tab restart button POSTed to a backend endpoint that (a)
checked the app's compose path from INSIDE the webui container, where
the host's containers root isn't mounted — so every restart failed with
'Compose file not found' — and (b) queued a raw 'docker compose restart'
that the host task processor would run as the manager user, which can't
talk to the rootless daemon anyway. Errors surfaced via a bare alert().
Per-service restart now follows the exact shape of the whole-app verbs:
- CLI: 'libreportal app restart <app> [service]' — the optional service
arg makes dockerRestartApp restart just that compose service, via
dockerCommandRun (right user in rootless mode) from the app dir on the
host, where the compose file actually lives. Service names validated
against compose-legal characters before touching a shell line.
- WebUI: the button dispatches a 'service_restart' task action through
the task router (mutations-via-tasks), runs in the background with the
standard task toast + link — no page switch — and failures use the
notification system instead of alert(). Because the task runs host-
side, restarting the WebUI's own libreportal-service now works too.
- Backend: the mutating restart endpoint and its now-unused helpers are
removed; service-routes.js is read-only surface (status + log tails).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Closes the gap behind the vpn-recreate bug: when the shared network is
recreated with a different /24, every app's stored static IP is left
outside it and adoptDockerSubnet only realigns CFG, not the apps.
- networkScanConflicts (network_conflicts.sh): read-only scan diffing each
active network_resources IP against docker's real subnet (via ipInSubnet).
Per-service routing-aware — skips gateway-routed services whose ipv4 is
commented out in the deployed compose, so gluetun apps don't false-positive.
Distinguishes 'daemon down' (benign) from 'network missing' (real).
- webuiSystemNetworkCheck (webui_system_network.sh): self-throttled generator
that writes frontend/data/system/network_status.json (modelled on
verify_status.json). Wired into webuiSystemUpdate AND run unconditionally
every ~60s from the task-processor poll (regen webui is mtime-gated and
would never fire on drift, which touches no source file).
- networkHealConflicts (network_heal.sh) + 'libreportal system network
check|heal [app]': the heal adopts docker's subnet in-process, then re-IPs
stranded apps with reset_network=ip (ports preserved), gluetun first.
Mutating path runs only through the task system (dual-mode, like update
apply); read-only check runs inline.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Foundations for network-drift healing:
- ipInSubnet(ip, cidr): prefix-aware CIDR membership (pure bash), so
stored IPs can be checked against docker's real subnet. Honours the
actual prefix, so a healthy /16-subnet + /24-ip-range install is not
mistaken for drift.
- dockerInstallApp now accepts reset_network="ip": re-roll the static IP
from the current subnet but PRESERVE published host ports (clears only
IP rows; LIBREPORTAL_RESET_IP_ONLY keeps port_allocate reusing existing
ports). This is the heal path — a subnet move strands the IP, not the
port, so we don't churn bookmarks/forwards/proxy upstreams. reset="true"
still re-rolls both.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
installDockerNetwork errored with 'network with name <x> already exists'
on re-runs: the requirement check sets DOCKER_NETWORK_SETUP_NEEDED=true
whenever 'docker network inspect' returns non-zero, but that also happens
when the rootless daemon socket isn't reachable yet — indistinguishable
from the network being genuinely absent. A prior install also leaves the
network behind, so the flag fires on every re-install.
Re-check existence right before creating and converge: if the network is
already there, leave it in place and adopt its real subnet into CFG rather
than erroring. This also stops the spurious subnet randomization (and the
resulting CFG drift) that ran before the doomed create.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The 'Installing System Requirements' step ran apt-get install with no
output until checkSuccess reported afterwards, so it looked frozen
while packages were being fetched. Print a notice up front.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A 4-lens adversarial security review of the Phase 2 applier raised 19 issues
and confirmed 17 after per-finding verification. All are trust-boundary (they
require the signing key), but several break the explicit "no code-exec, always
reversible, nothing-silent" contract, so all 17 are fixed:
Trust path — fail CLOSED, never misreport:
- lpFetchIndex now surfaces the real signature state (LP_INDEX_SIGSTATE);
artifactApply REFUSES to mutate unless the index is actually verified, and
_artifactFetchPayload refuses an unsigned payload. The read path still
tolerates dev/unsigned but now says "UNSIGNED" instead of "Signed + verified".
- valid_until and index_serial are now MANDATORY + numeric in lpFetchIndex
(missing = refuse) — closes the anti-withholding / anti-rollback fail-opens.
Injection / code-exec (defense in depth even for a signed payload):
- runFileWrite rootless branch no longer builds a `bash -c` shell string with the
destination interpolated — it uses the argv form (like runFileOp), so a path
with a quote can't inject a command as the install user. (shared-helper fix)
- op paths must match a safe-filename charset (no quotes/$/backtick/;/newline);
set-config-key values and set-compose-image refs are charset-guarded too.
- content_b64 is validated as real base64 at precheck.
Reversibility / honest failure:
- dockerComposeUp now returns the real compose exit status (it always returned 0,
so the updater's rollback gate AND the apply's start-failure detection were
fail-open). (shared-helper fix)
- set-config-key undo captures the WHOLE config file (lossless) instead of a
lossy re-parsed scalar; edit-only (rejects an absent key).
- _artifactReplayUndoFile returns non-zero if any inverse op fails; auto-rollback
and revert now record "rollback-incomplete"/"revert-incomplete" + isError
instead of falsely claiming success, and revert keeps the record for retry.
- applied-record write failure is checked — apply rolls back rather than leave an
un-revertable change. System-scope regen failure is no longer swallowed.
- Writes are path-aware (configs/ -> runInstallWrite, container tree ->
runFileWrite) so system-scope hotfixes write/restore correctly.
- Checked lazy-sourcing surfaces a clear error instead of a bare exit 127.
Unit-tested 35/35 (adds: command-sub value rejection, bad image-ref, invalid
base64, quote/metachar path-injection rejection, replay-failure reporting).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The honest-checkSuccess + masking fixes immediately surfaced a real masked
failure in error_report.log: updateDockerSudoPassword (run every system scan
from start_scan.sh) does 'sudo passwd $sudo_user_name', but Model A's scoped
sudoers grants only LP_HELPERS/LP_SYSTEM + run-as-install-user — not passwd.
So at runtime (manager, non-root) it failed exit 1 every scan, masked until now.
The password is set at install (root, chpasswd) and admin login is key-based,
so the runtime re-sync is legacy + impossible under de-sudo: guard it to skip
unless EUID 0. (Validates the surfacing mechanism working as intended.)
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
'local result=$(cmd)' resets $? to 0 (the local builtin's own exit), so the
following checkSuccess always saw success regardless of cmd's real exit — the
mechanism that masked the de-sudo write failures. Split declaration from
assignment ('local result; result=$(cmd)') across all 235 active-code sites
(84 files) so the command's exit reaches checkSuccess. No behaviour change
beyond $? now being accurate (no set -e in runtime code; multi-line
assignments transform safely).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Two more cases of the manager writing directly into the container-owned
/libreportal-containers tree (same class as the regen-poll stamp), both masked
by a '✓ Success' that printed anyway:
- Password replacers (config/password/*): used 'runInstallOp sed -i' (manager)
on app configs copied into the container tree, so sed -i EACCES'd its temp
file and the substitution silently failed — the adguard.config 'couldn't open
temporary file', leaving the literal RANDOMIZEDPASSWORD placeholder. Added
runCfgOp (picks runFileOp vs runInstallOp by the target file's location) and
routed every $file grep/sed/awk through it: password, username, hex, vapid,
appkey, and bcrypt.
- Updater generator (webui_updater_scan): 'runFileOp cp <manager-tmp>' can't
read the manager's 0600 mktemp as the container user, so it fell through to a
manager 'cp' that EACCES'd on the container-owned out_dir. Switched the three
writes to 'runFileWrite < tmp' (manager shell reads the tmp; container user
tees the write).
Both deploy via the normal quick path (relocatable scripts) — no footprint bump,
no reinstall.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Two small uninstall-output tweaks.
1. dockerComposeDownRemove now ALWAYS calls dockerRemoveApp (the
`docker ps -aqf name=…` → stop + rm sweep) as a fallback, even when
the compose-down step is skipped because the app dir is missing.
Before, a partial prior uninstall (compose file gone but containers
still running) produced "App directory not found. Skipping container
shutdown." and then proceeded as if the uninstall were complete —
leaving the actual containers running. The name-based sweep also
runs after a successful compose-down to catch anything compose
wouldn't pick up (renamed services, orphans from earlier failures).
While here: the OS_TYPE gate (only Ubuntu/Debian) is gone too —
`docker compose down` works on any OS with docker, and gating it
meant Arch/etc. users got NO compose teardown at all.
2. The step-2 header "Keeping Docker images (pass --delete-images to
remove)" trimmed to just "Keeping Docker images". The `isNotice`
line below already explains the reuse-on-reinstall behaviour; the
CLI-flag hint reads as noise in the WebUI task log where users
can't act on it anyway. CLI users can still pass --delete-images
(cli_app_commands.sh wires it as before) or tick the WebUI's
"Also delete docker image" checkbox.
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerDeleteData (uninstall) and the wipe-before-restore step in
restoreAppStart both did `runFileOp rm -rf $containers_dir$app_name`,
which runs as $CFG_DOCKER_INSTALL_USER (dockerinstall, uid 1002 on
rootless). That user owns app-template files but CANNOT remove
container sub-UID dirs created by the daemon's userns mapping —
postgres data at uid 232070, nextcloud html at uid 33, etc. The rm
therefore silently failed with
rm: cannot remove '/libreportal-containers/invidious/postgresdata':
Permission denied
while still reporting "<app> successfully uninstalled" — leaving the
sub-UID directory tree on disk to confuse the next install and leak
storage.
Fix: route the wipe through a new `app-data-remove` action in the
root-owned libreportal-ownership helper. Root can rm sub-UID files
unconditionally. The helper validates the app name (alphanumeric +
. _ -, no traversal), refuses the WebUI's own slot (libreportal), and
is idempotent when the dir is already gone.
Two callers updated:
- scripts/docker/app/uninstall/delete_data.sh
- scripts/restore/restore_app_start.sh
The helper itself ships root-owned at /usr/local/lib/libreportal/, so a
fresh install or release upgrade is needed to pick up the new action.
Bumped init.sh footprint_version 2 → 3 so the runtime updater
prompts a root re-install on the next release.
Signed-off-by: librelad <librelad@digitalangels.vip>
The 31 containers/<app>/<app>.sh files each defined install<App>() with
the SAME 10-step sequence — ~4,000 lines of duplicated boilerplate.
Replaces all that with one generic driver + hook surface.
scripts/app/install/app_install.sh:
installApp <slug> [config_variables]
— Dispatches on $<slug> (c/u/s/r/i) the same way the per-app .sh
files did. Same convention; dockerInstallApp's existing
`declare $app=i` callsite needs no change.
— Runs the standard sequence: dockerConfigSetupToContainer →
dockerComposeSetupFile → optional .env copy → fixPermissions →
dockerComposeUpdateAndStartApp → standard post-install steps
(appUpdateSpecifics, setupHeadscale, databaseInstallApp,
webuiContainerSetup, monitoring registration) → final message.
— Hooks (all declare-f-gated, silent no-op when absent):
<slug>_install_pre / _post_setup / _post_compose / _post_start
<slug>_install_message_data (echoes extra args for menu)
<slug>_install_post
<slug>_uninstall_pre / _post
<slug>_stop_post
<slug>_restart_post
Hooks live in containers/<app>/tools/<app>_tools.sh (auto-sourced
per the modular-per-app-tools convention).
function_install_app.sh:
When no install<App>() function exists, fall through to
`installApp <app_name>` instead of erroring. So an app with no .sh
at all becomes a zero-byte addition — drop in <app>.config +
docker-compose.yml + <app>.svg, done.
containers/linkding/linkding.sh:
Deleted (canary). Linkding's body was 100% standard sequence;
fallback handles it identically. Smoke-tested with stubbed helpers
— dispatcher fires, generic runs full flow, monitoring integration
+ final-message hook plumbing all intact.
Wave B (next): delete the .sh for every other 'pure-boilerplate' app
(~15 candidates per the survey). Wave C: extract custom logic from
the 7 fat apps into hooks before deleting their .sh.
Signed-off-by: librelad <librelad@digitalangels.vip>
The Debian-shipped passt AppArmor profile (/etc/apparmor.d/usr.bin.passt)
denies the accesses pasta needs to plumb rootlesskit's netns:
- ptrace_read on the rootlesskit child to enter its user namespace
- read /run/user/<uid>/dockerd-rootless/netns (the netns file)
- read /proc/<pid>/net/{tcp,tcp6,udp,udp6} for implicit port forwarding
Without these the rootless docker daemon fails with:
pasta failed with exit code 1:
Couldn't open user namespace /proc/<pid>/ns/user: Permission denied
scripts/docker/install/rootless/rootless_apparmor.sh:
New installRootlessApparmorForPasta() — idempotent fixup.
1. Adds `include if exists <local/usr.bin.passt>` to the main profile
(one line; re-adding is a no-op via grep).
2. Writes /etc/apparmor.d/local/usr.bin.passt with the four rules
pasta needs. The /local/ pattern is the standard Debian AppArmor
hook for site-managed overrides — survives `apt upgrade passt`
because it's outside the package's managed paths.
3. Reloads via apparmor_parser -r.
Called from installDockerRootless after the override.conf write, gated
on $rootless_net == pasta. slirp4netns installs skip it.
This box was already manually patched while debugging the pasta swap —
the installer-side change makes it idempotent across reinstalls and
applies the same fix on any other host that installs rootless docker
with pasta as the net driver.
Signed-off-by: librelad <librelad@digitalangels.vip>
CrowdSec's host-side install (the agent + nftables bouncer the LibrePortal
Traefik plugin talks to) had stayed on blanket sudo throughout the rootless +
de-sudo hardening: `sudo apt-get install crowdsec`, `curl | sudo bash`,
`sudo sed -i /etc/crowdsec/config.yaml`, `sudo touch + sudo chmod /var/log/
crowdsec*.log`, `echo $key | sudo tee /etc/crowdsec/traefik_bouncer.key`,
plus `sudo cscli capi register / console enroll / bouncers add`. None of
those are in the scoped LP_HELPERS / LP_SYSTEM sudoers grant the manager
now holds, so any user who enabled crowdsec would have hit hard sudo
failures on every privileged step.
Follow the libreportal-appcfg / libreportal-bininstall pattern: one new
root-owned helper at /usr/local/lib/libreportal/libreportal-crowdsec
that does every privileged op behind a fixed action vocabulary with strict
argument validation. The manager calls in via runCrowdsec — the scoped
sudoers grants exactly one binary, the same trust boundary the other
helpers rely on.
Actions:
install apt repo + agent + firewall-bouncer + enable +
crowdsecurity/{linux,sshd} collections + reload
(idempotent — skips parts already in place)
services <verb> enable | disable | restart
capi <verb> register | unregister | status
console <verb> enroll <token> | disenroll | status
token format strictly validated
bouncer-traefik-init cscli register + write the manager-owned key file
atomically (returns EXISTS or GENERATED:<key>)
bouncer-priority bouncer yaml nftables priority → -100
(moved from libreportal-appcfg; one helper for
every crowdsec root op)
bind-lapi flip listen_uri to 0.0.0.0:8080 in config.yaml
prometheus <on…|off> flip the prometheus block (validated addr/port)
touch-host-logs create + chmod 0644 /var/log/crowdsec*.log so the
libreportal container can tail them
Wired in via:
- new sudoers Cmnd_Alias entry for the helper in LP_HELPERS
- new helper baked alongside the others by initRootHelpers
(replaces __SYSTEM_DIR__ / __CONTAINERS_DIR__ / __MANAGER__ at
install, with safe runtime fallbacks if unbaked)
- new runCrowdsec dispatch in scripts/docker/command/run_privileged.sh
containers/crowdsec/scripts/crowdsec_install_host.sh now drives the whole
flow through runCrowdsec — every `sudo …` is gone, the compose-toggle sed
uses runFileOp, and the security_crowdsec CFG mirror uses runInstallOp
(configs/ is manager-owned). Net: install script shrinks ~80 lines while
gaining a single auditable trust boundary. crowdsec_fix_priority.sh swung
over to runCrowdsec bouncer-priority too — the appcfg crowdsec_priority
action drops out cleanly.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
runBackupOp dropped privileges to $docker_install_user with `sudo -E`,
which preserves the CALLER's environment — including HOME. The caller is
the manager (libreportal), so restic-running-as-dockerinstall ended up
with HOME=/home/libreportal and tried to mkdir
`/home/libreportal/.cache/restic` for its cache. dockerinstall can't
write into libreportal's home, so every backup ran with:
unable to open cache: mkdir /home/libreportal/.cache/restic: permission denied
twice (once in backup, once in the verify-via-scratch-restore step), with
restic falling back to a no-cache run that's a few × slower than it
should be.
Add `-H` (sudo's "reset HOME to target user's home"). Now restic sees
HOME=/home/dockerinstall, creates ~/.cache/restic there (dockerinstall
owns its own home, no help needed), and the warning is gone. Confirmed
live: a `backup app create linkding` round-trip is silent on cache, and
the dir lands at /home/dockerinstall/.cache/restic, mode 0700, correctly
owned.
All restic/borg/kopia calls funnel through runBackupOp, so this single
character fix covers every backup-tool invocation.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The migrate/ helpers were either uncallable or no-ops:
- migrateAppsToNewNetwork + updateComposeFileNetwork: never called from
anywhere. The intended sed-on-compose subnet rewrite would also have
fought the tag system / network_resources DB.
- checkAppNetworkCompatibility: called from updateDockerNetworkConfig as
a gate, but never explicitly returns, so it's effectively always-true
and both branches do the same work. Pure noise.
- getInstalledApps: only used by the above.
- updateDockerNetworkConfig: collapses to a 2-line 'CFG := docker's
reported subnet' adoption — inlined into check_docker_network.sh as
adoptDockerSubnet(), which is what it actually does.
The legitimate 'subnet changed, refresh apps' path is already covered by
the idempotent per-app reinstall (dockerInstallApp ... reset_network=true
→ clears DB allocations → installer re-runs → ipUpdateComposeTags picks
fresh IPs from the current CFG_NETWORK_SUBNET). Migration (infrastructure
regen) vs restore (data) stays clean: reinstall regenerates compose+IPs,
restore lays data on top. No new pathway needed.
Files dropped:
scripts/docker/network/migrate/migrate_apps_to_new_network.sh
scripts/docker/network/migrate/migrate_check_app_network_compatibility.sh
scripts/docker/network/migrate/migrate_get_installed_apps.sh
scripts/docker/network/migrate/migrate_update_compose_file_network.sh
scripts/docker/network/migrate/migrate_update_docker_network_config.sh
Plus the now-empty migrate/ subdir; files_docker.sh regenerated to drop
the references.
Signed-off-by: librelad <librelad@digitalangels.vip>
If it's gluetun code, it lives with gluetun. Both functions in
scripts/config/tags/processors/tags_processor_network_mode.sh manipulate gluetun
markers / gluetun's compose, so move them into containers/gluetun/scripts/
gluetun_network.sh and rename to the per-app-hook convention:
tagsProcessorNetworkMode -> appNetworkApplyMode_gluetun
tagsProcessorGluetunForwardedPorts -> appNetworkRegisterPorts_gluetun
Central call sites are now provider-agnostic — no "gluetun" literal anywhere:
- docker_config_setup_data.sh: an app routing via CFG_<APP>_NETWORK=<provider>
triggers `appNetworkApplyMode_<provider>` + `appNetworkRegisterPorts_<provider>`
via declare -F, so any future gateway provider plugs in with no engine edits.
- uninstall_app.sh: loops every `appNetworkRegisterPorts_*` hook (each self-skips
when its provider isn't installed), so removing a routed app refreshes the
right provider with no provider name in central code.
Delete tags_processor_network_mode.sh; regenerate arrays. Verified with stubs:
default mode no-ops, gluetun-routed app fires both hooks, gluetun itself is
skipped, unknown provider is silently no-op, uninstall loop calls registerPorts.
Drive-by cleanup: 9 stale "${X_scripts[@]}" array references in app_files.sh /
cli_files.sh (gluetun + headscale from this session's moves, plus 7 pre-existing:
command/ssl/swapfile/ufw/ufwd/user — all from older refactors that left them
behind). Each expanded to nothing at runtime (harmless), but they're dead
misleading refs. Cleaned both files; every remaining array ref now points to a
real files_*.sh.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerCheckAllowedInstall was a one-app `case` whose only active caller was the
wireguard app itself — so inline its check (abort if a host WireGuard exists at
/etc/wireguard/params, which would collide on the wg kernel module + UDP 51820)
directly into containers/wireguard/wireguard.sh and delete
scripts/docker/app/checks/allowed_install.sh.
The protection is unchanged; wireguard is now fully self-contained and the last
app name leaves central install code. Regenerated arrays. (The only remaining
dockerCheckAllowedInstall references are in scripts/unused/ — retired apps,
never sourced.)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Each app now carries everything under containers/<app>/: Tools-tab actions in
tools/ (declaration <app>.tools.json + function <app>_<tool_id>.sh) and logic
helpers in scripts/ (e.g. <app>_auth.sh). The container scan live-sources every
.sh under the app (maxdepth 3, prunes only resources/) and webui_tools.sh
auto-merges the .tools.json, so an app is a true drop-in — no central edit, no
array regen.
- Empty the central webui_tools.sh heredoc; all 34 tools across 11 apps now
come from per-app declarations (verified byte-identical to the old output).
- Retire the orphaned mattermost tool scripts to scripts/unused (there is no
containers/mattermost; its install fn already lived in unused).
- Update the dispatch comment/error path, the auth-adapter doc, and
DEVELOPMENT.md to the new convention.
- Regenerate static arrays (files_app.sh no longer lists app/containers/*).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Introduce scripts/source/paths.sh as the canonical path resolver for three
independently-relocatable roots:
LP_SYSTEM_DIR manager-owned control plane (configs/logs/install/db/ssl/ssh/migrate)
LP_CONTAINERS_DIR container-user-owned live app data
LP_BACKUPS_DIR container-user-owned backup repos (own mount-able)
Roots come from the environment when set (install bakes them; CLI/app inherit
from init.sh), else default to /libreportal-*. A transitional compat default
keeps EXISTING installs (legacy single /docker tree, by config marker) on /docker
until a deliberate reinstall, so deploying this never strands a running box.
- init.sh derives the same vars inline (self-contained for the bare /root/init.sh
reinstall case); paths.sh mirrors it for the standalone task/check processors,
which now self-locate their scripts dir and source it.
- Replace functional /docker literals with the derived vars across runtime,
install, backup, crontab, crowdsec/restic, headscale, and reinstall paths;
clean the inert '== /docker/containers/*' guard fallbacks to the variable form.
- backend: CONTAINERS_DIR now from LP_CONTAINERS_DIR (compose env, filled at
generation via a new CONTAINERS_DIR_TAG), legacy-safe default for un-recreated
containers.
- backup default path falls back to the backups root; exclude paths.sh from the
sourced-file arrays (bootstrap file, sourced explicitly).
The CLI-wrapper heredoc + root helpers still reference /docker; those get baked
in phase 3. No layout/ownership change yet (phase 2).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Reverts the 2>/dev/null band-aids and fixes the root cause. The
manager-run install boot scans app configs under /docker/containers AS
the container user (runFileOp). But init.sh's initFolders creates that
dir manager-owned, and the handover to the container user happened later
(start_preinstall), AFTER the boot scans — so the scans ran as the
container user against a dir it didn't own yet: "find:
'/docker/containers/': Permission denied" (cosmetic; the dir is empty
that early, but it's the wrong ownership at the wrong time).
Add initContainerLayer() to init.sh's root phase (after initGIT +
initUpdateConfigs, before the manager-run handoff): rootless-only, it
creates the docker-install user if missing and chowns /docker/containers
to it (751). The later rootless setup is now idempotent — it finds the
user existing and just (re)asserts its password + daemon config (moved
updateDockerInstallPassword out of the create-only branch). Rooted is
unaffected (containers stay manager-owned, which the manager reads).
Result: by the time the boot scans run, /docker/containers is owned by
the user doing the scanning — no permission error, nothing suppressed.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
On a fresh install the requirement checks run before the things they
probe exist, leaking raw command stderr:
- check_install_type.sh: `$( (id -u "$user") )` printed
"id: 'dockerinstall': no such user" to the terminal AND — since id's
error goes to stderr, not the captured stdout — the next line's
`[[ "$ISUSER" == *"no such user"* ]]` could never match, so the
rootless-user-absent branch was dead. Add `2>&1` (matching siblings on
lines 25/31): no leak, and the check now works.
- grep on $sysctl (the rootless marker conf, absent until rootless is set
up) printed "grep: /etc/sysctl.d/99-libreportal-rootless.conf: No such
file or directory". Add -s to the four $sysctl greps
(check_docker_rootless, rootless_start_setup, rootless_docker x2);
"marker absent" is still detected (non-zero exit), just without the
file-not-found message.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The two docker-type-switcher finds run mid-switch, BEFORE
reconcileDockerOwnership, so containers/ is still owned by the OLD mode's
container user while CFG_DOCKER_INSTALL_TYPE is already the target. A
plain runFileOp resolves to the target user, which can't list the
old-mode-owned (751) dir under rootless — so enumerate as the old-mode
owner instead:
- switchMigrateBackupApps: move the find inside the existing
old_mode/resolveDockerInstallUser window (runFileOp now resolves to the
old owner). It previously ran as the manager and silently enumerated
nothing under rootless, so no app got backed up before the switch.
- dockerSwitcherUpdateContainersToDockerType: take old_mode as an arg,
flip CFG to it only for the find (restore before the per-app socket
scan + restart, which need the new daemon). Callers in swap_docker_type
pass $docker_type. The two former rooted/rootless branches were
byte-identical and are collapsed.
NOTE: the full rooted<->rootless switch round-trip is still unvalidated
on the VM (needs a stateful app + an enabled backup location); this fixes
the container enumeration, not yet the end-to-end migration.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Bare `find "$containers_dir"` runs as the manager, but under rootless
containers/ is dockerinstall-owned 751 (traversable, not list-readable by
the manager) -> "find: /docker/containers/: Permission denied". For the
app-log generator that was cosmetic; for dockerComposeUpAllApps /
dockerComposeDownAllApps it silently enumerates nothing so no apps come
up/down. Route these through runFileOp find (dockerinstall in rootless,
manager in rooted — correct in both). The two docker-type switcher finds
are deliberately left: mid-switch the at-rest container owner can differ
from the target-mode user runFileOp resolves to, so they need mode-aware
handling rather than a blind swap.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Organise the system footprint outside /docker:
- All LibrePortal executables now live together in /usr/local/lib/libreportal/
(root:root): the 7 root helpers AND the CLI wrapper. /usr/local/bin/libreportal
becomes a symlink onto $PATH. run_privileged._runRootHelper, init.sh
(initRootHelpers + scoped-sudoers Cmnd_Alias + command setup) all point there.
The wrapper is now root-owned too (manager can't tamper with its entrypoint).
- Fix a real bug: rootless sysctl settings were written to /etc/sysctl/99-custom.conf,
a dir does NOT read, so net.ipv4.ip_unprivileged_port_start /
kernel.unprivileged_userns_clone never persisted across reboot. Moved to
/etc/sysctl.d/99-libreportal-rootless.conf (the existing
reload now actually applies them). Consistent libreportal* naming.
- Drop dead fqdn_file=/root/libreportal-fqdn.txt global (never used).
- Add FOOTPRINT.md: a manifest of every file LibrePortal places outside /docker
(doubles as an uninstall checklist).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Bring the remaining deferred subsystems under the scoped sudoers, and drop
the one that's redundant.
Backup engines + app configs -> root-owned helpers (same pattern as
ownership/dns/ssh/socket/svc):
- scripts/system/libreportal-bininstall: install <restic|kopia> — does the
whole pkg-manager/signed-download install itself for a fixed, validated
engine name (no blanket sudo apt-get/install). restic_install/kopia_install
call it.
- scripts/system/libreportal-appcfg: {adguard-auth <user> <bcrypt>|
crowdsec-priority|owncloud-config <public> <host> <ip> <public_ip>} —
faithful ports of the AdGuard yaml / CrowdSec bouncer / ownCloud config.php
rewrites, fixed paths + validated args. adguard_auth/crowdsec_fix_priority/
owncloud_setup_config call it.
- run_privileged: runBinInstall / runAppCfg; init.sh installs + allowlists both.
Retire standalone (host-level) WireGuard — it's a duplicate of the
containerized containers/wireguard app (+ headscale mesh), its slirp4netns
speed rationale is largely moot with a better rootless net backend / typical
WAN-bound throughput, and it was the heaviest host-root subsystem (apt +
sysctl + iptables + /etc/wireguard), the worst fit for the rootless/
least-privilege direction:
- moved scripts/wireguard/ + manage_wireguard.sh + check_wireguard.sh to
scripts/unused/; dropped the install-path call, the Tools menu 'w' entry,
and the requirement check; removed the half-built libreportal-wg helper.
- generate_arrays.sh now also skips system/ (root-owned helpers, never
sourced); arrays regenerated (files_wireguard.sh pruned).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The scoped sudoers grants the manager (root) and (dockerinstall) but NOT
(itself), so the many 'sudo -u $sudo_user_name <cmd>' calls (crontab,
git/update, reinstall, swapfile, …) failed with 'a password is required'
once per CLI command. runAsManager runs the command plainly when already
the manager (the runtime case) and only sudo -u's when root (install
time), so it's correct in both contexts and needs no sudoers self-grant.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Move the last runtime-critical root file-primitive subsystems behind
root-owned helpers so the type switcher + task service work under a scoped
sudoers:
- scripts/system/libreportal-socket: {rootless|rooted} {on|off} chmod of
the docker sockets (paths computed from config, not caller-supplied;
exit 3 = absent so the *_found flags come from its exit code)
- scripts/system/libreportal-svc: GENERATES + installs the systemd unit
from config (mode/uid/baked manager) — never accepts unit content from
the caller (arbitrary unit = root). Idempotent install/enable/restart.
- ownership helper: add db-own + app-file <app> <relpath> actions
- run_privileged: runSocket / runSvc
- set_socket_permissions -> runSocket; webui_install_systemd -> runSvc
(+ crontab cleanup runs as the manager directly, no sudo -u self)
- before_start: db chown -> runOwnership db-own; traefik cert/yml ->
runOwnership app-file (retires updateFileOwnership/changeRootOwnedFile)
- init.sh installs all five helpers
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Two more runtime root file-primitive subsystems moved behind self-
validating root-owned helpers so the scoped sudoers needn't grant blanket
sudo sed/tee/cp on /etc (which is root-equivalent — sudo arg wildcards
match across '/', so even path-scoped entries are bypassable):
- scripts/system/libreportal-dns: {clear|add <ip>} — edits /etc/resolv.conf
only, validates the IP argument
- scripts/system/libreportal-ssh-access: authorized_keys + sshd
PasswordAuthentication management, with the lockout guards moved INTO the
helper (the trust boundary) so a compromised manager can't bypass them
- run_privileged: _runRootHelper dispatcher + runResolv / runSshAccess
(runOwnership now uses it too)
- init.sh: initRootHelpers installs all three helpers root:root 0755 with
the manager name baked in
- setup_dns -> runResolv (+ ping de-sudo'd, works unprivileged); host_access
+ webui_ssh_access -> runSshAccess
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Under Model A the runtime runs as the manager, so establishing the
/docker ownership model needs root. Granting the manager a blanket
'sudo chown'/'sudo chmod' in the scoped sudoers would be root-equivalent
(chown /etc/sudoers, ...). Introduce a self-contained, root-owned helper
that performs only a FIXED set of reconciles on FIXED LibrePortal paths,
with owners derived from config + a baked manager name (never the caller)
and a strictly-validated app-name argument.
- scripts/system/libreportal-ownership: the helper (actions: reconcile,
traversal, containers-top, app-perms, webui, taskdir, app-data-nobody)
- run_privileged: runOwnership wrapper (sudo the installed helper; run the
bundled copy directly when already root mid-install)
- init.sh: installOwnershipHelper bakes the manager name and installs it
root:root 0755 to /usr/local/sbin (manager can't modify it)
- libreportal_folders/app_folder/app_update_specifics/task processor:
delegate the ownership chowns to runOwnership instead of runSystem chown
This removes chown/chmod-on-/docker from the runtime sudo surface, a
prerequisite for a non-root-equivalent scoped sudoers.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- docker_run: in rooted mode run docker AS the manager via the docker
group (no sudo); the type=='sudo' branch was unreachable dead code
- 8 db helpers: fix 'command -v sudo sqlite3' guard to 'command -v
sqlite3' (bodies already query via runInstallOp)
- restic/kopia single-file dump: write target_file via runBackupOp tee
(as the backup user, matching the snapshot-restore path) instead of
root tee
- adguard auth: root-owned scratch via runSystem mktemp
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The borg/restic/kopia engines all dropped to the dedicated backup user
via scattered 'sudo -E -u $docker_install_user'. Centralize that into a
single runBackupOp helper so the backup subsystem has one audit point and
the scoped sudoers needs only the (dockerinstall) drop rule.
Also:
- owncloud config heredoc tees -> runSystem (container-UID file)
- webui_display_logins: fix the broken 'command -v sudo sqlite3' guard
to 'command -v sqlite3' (body already runs sqlite3 via runInstallOp)
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
- copy_build_context: rsync/cp/rm -> runFileOp (writes the deployed tree AS the
container owner with --no-owner); drop the now-redundant runSystem chown.
- setup_lock: .setup_complete is in the docker-install-owned frontend/data ->
runFileOp touch/chmod/rm (drop the chown).
- tags_processor_docker_installation 'user:' enable + update_compose_yml
jail.local -> runFileOp (deployed compose/config under containers).
- crontab_clear: clear the manager's own crontab via runInstallOp.
- reinstall: cp init.sh to /root -> runSystem (genuine root path).
- create_successful_run_file: drop the pointless sudo echo -> runInstallWrite to
/docker/run.txt.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Model A prototype (run start.sh AS the manager, escalate only via helpers):
- check_root.sh: accept the manager user, not root-only (init.sh keeps its own
install-time root check).
- init.sh: guard the top-level root-check + installer entrypoint with
BASH_SOURCE!=$0 so it runs ONLY when init.sh is executed directly; when
start.sh sources it as the manager the entrypoint (and its root check) no
longer fires.
Also: convert bare daemon-touching 'docker' calls (no helper -> hit the
nonexistent /var/run socket in rootless) to runFileOp docker across
app_status, app_health_*, network_prune, ip_is_available, check_docker_network,
backup_db (db dumps) and crontab_check_processor. cd&&compose rooted-branches
and 'docker compose --version' checks left as-is (rooted-only / no daemon).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Container-plane docker now routes through the mode-aware helpers instead of
sudo: simple calls (exec/ps/run/build/images/inspect/port/logs across ~15
app/check scripts) -> runFileOp docker (rootless socket as the install user;
rooted via the docker group). The cd && docker compose paths drop the sudo on
the rooted branch (the rootless branch already used dockerCommandRunInstallUser
-- byte-identical now, manager-ready later); gluetun, which had no rootless
branch, now uses dockerCommandRun so force-recreate works in both modes.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The apps SQLite DB ($docker_dir/$db_file) is owned by the manager user, so
read/write it AS the manager via runInstallOp instead of sudo (root). 48 call
sites across 28 scripts. In rooted this drops root->manager (correct owner);
in rootless it's the manager too (using runFileOp/dockerinstall here was the
'unable to open database' bug). The broken 'command -v sudo sqlite3' check
lines are left untouched (separate pre-existing issue).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Foundation for a scoped sudoers: route every genuine system-admin command
(systemctl/ufw/ufw-docker/nft/apt/apt-get/pacman/sysctl/useradd/usermod/
service/wg/wg-quick/cscli/loginctl) through runSystem instead of raw sudo
across 28 active scripts. runSystem is 'sudo "$@"' so this is byte-identical
in every mode (safe on live installs) — it just collects all real-root use at
one chokepoint that will define the eventual /etc/sudoers.d allowlist.
Also: revert a crowdsec advice message the sweep wrongly rewrote (the admin
types sudo, not runSystem), and give crontab_check_processor.sh the same
startup bootstrap as the task processor — it runs standalone via cron and
already used runFileOp/runFileWrite (undefined there), so it was silently
broken; now it sources the helpers + docker-type config.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
installLibrePortalWebUITaskService only wrote the unit if it didn't already
exist, so env/User/mode changes never reached an existing install and a
docker-type switch couldn't update the service. Make it converge: compute the
desired unit for the current mode and only rewrite + daemon-reload + restart
when it actually differs (otherwise just ensure enabled+running, no restart, so
routine re-runs don't bounce the processor and kill in-flight tasks). The
docker-type switcher now calls this idempotent setup (replacing the one-shot
restart helper), so a swap updates the env AND restarts in one step.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The task processor reads CFG_DOCKER_INSTALL_TYPE once at startup to decide how
runFileOp writes into the task dir (rootless -> as the docker install user,
rooted -> as the manager). After a rooted<->rootless swap a running instance
keeps the old mode and writes task files wrong. Add
restartLibrePortalWebUITaskService and call it at the end of both switch
branches so the processor re-sources the new mode. The switch is a CLI
one-shot, not a processor task, so the restart won't interrupt it.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerCopyBuildContext rsync'd the install template into the container dir
with -a, which preserves owner/group — so the deployed WebUI tree (frontend/
included) inherited the repo clone's owner (the human user, uid ~1000) on
every install. The trailing chown used the $docker_install_user global, which
is stale/empty in this context, so it silently no-op'd and uid 1000 survived
(visible as frontend/ owned by 1000 with the template's mtime).
Add --no-owner --no-group so the copy doesn't carry source ownership, and
chown via the config-authoritative dockerContainerOwner (rooted -> manager,
rootless -> docker install user) through runSystem. The deployed tree now
lands owned by the mode's container owner.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerCommandRunInstallUser sudo's to the unprivileged docker install user but
inherited the caller's cwd. At install time the caller is root in /root, which
that user can't enter, so cwd-sensitive tools failed — e.g. 'find: Failed to
change directory: /root' / 'Failed to restore initial working directory'
during the app scan (the scan still worked via the absolute start path, but
the errors are noise and could bite other commands). Add env --chdir to the
install user's HOME for both the argv and shell exec paths so every runFileOp
runs from a directory the user can access.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
Switching rooted<->rootless re-maps every container's on-disk UIDs (rootless
offsets them by the subuid base), so a stateful app's data no longer lines up
in the new mode and a chown can't carry it. The portable carry is backup
(old mode) -> switch -> restore (new mode): restoreAppStart wipes and re-lays
each tree and re-owns it to the new mode's install user, which is exactly the
remap needed.
Wire that into dockerSwitcherSwap:
- switchMigrateBackupApps <old_mode>: before the switch, back up every
installed app except libreportal (reconcile already carries the control
plane). CFG_DOCKER_INSTALL_TYPE is already the target mode by the time the
switcher runs, so force it (and the resolved install user) back to the old
mode for the backups, else backupAppStart would talk to the not-yet-running
new daemon. Any backup failure aborts the switch before the daemon is
touched (nothing changed). No backup location enabled -> skip and keep the
manual warning.
- switchMigrateRestoreApps: after the new daemon is up, restore each captured
app best-effort, re-resolving the install user first so data is owned
correctly; failures are reported per app rather than blocking.
Subject to the existing backup-completeness limitation (restic-as-libreportal
can't read files owned by other UIDs unless the app declares container-side
file capture) — same caveat as the manual procedure this automates.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The switch-to-rootless branch passed the literal 'root' to mode-aware
helpers whose vocabulary is 'rooted'/'rootless' (matching
CFG_DOCKER_INSTALL_TYPE). Two were real no-ops: dockerComposeDownAllApps
root never matched dockerComposeDown's 'rooted' check (old rooted apps were
never composed-down before the switch), and dockerServiceStop root never
matched dockerServiceStop's 'rooted' check (the old rooted docker service
was never stopped/disabled). dockerServiceStart root was harmless only
because that function ignores its arg and reads CFG. Both no-ops were
silent until dockerComposeDown started reporting unknown modes. Align all
three to 'rooted'.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
The single long isNotice was hard to read in both source and terminal
output. Break it into three lines (re-map warning / backup-restore guidance
/ app-data note).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
dockerComposeDown printed the 'Docker Compose down <app>' header then could
fall through silently: when the effective install type (passed type arg or
CFG_DOCKER_INSTALL_TYPE) was empty/unrecognised no branch ran, and on a
non-Ubuntu/Debian OS the whole block was skipped. Collapse the duplicated
type=='' vs type!='' branches into one mode fallback and add notices for the
unknown-mode and unsupported-OS cases so the header always has a result line.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
Signed-off-by: librelad <librelad@digitalangels.vip>
